// Create a sandbox
const res = await fetch('https://sandbox.developersdigest.tech/api/v1/sandboxes', {
method: 'POST',
headers: {
'Authorization': 'Bearer YOUR_API_KEY',
'Content-Type': 'application/json',
},
});
const sandbox = await res.json();
// Execute a command
const exec = await fetch(
`https://sandbox.developersdigest.tech/api/v1/sandboxes/${sandbox.id}/exec`,
{
method: 'POST',
headers: {
'Authorization': 'Bearer YOUR_API_KEY',
'Content-Type': 'application/json',
},
body: JSON.stringify({ command: 'pip install pandas && python analysis.py' }),
}
);
const result = await exec.json();
console.log(result.stdout);
// Kill the sandbox when done
await fetch(
`https://sandbox.developersdigest.tech/api/v1/sandboxes/${sandbox.id}`,
{ method: 'DELETE', headers: { 'Authorization': 'Bearer YOUR_API_KEY' } }
);Try it live
No signup required. Python, Node.js, and Bash — running in a real Linux sandbox with hardware isolation.
Features
Every sandbox is a Firecracker microVM with its own Linux kernel. No shared kernel. No container escapes.
Firecracker snapshot restore. Your sandbox is ready before your LLM finishes generating the next token.
Each sandbox is a Firecracker microVM with its own Linux kernel. The same tech that powers AWS Lambda.
Real-time stdout/stderr via SSE. Watch your agent's code execute live with streaming callbacks.
Run any language, install any package, start any process. Root access to your own Linux machine.
Snapshot full VM state to disk. Resume later exactly where you left off. Pay nothing while paused.
Read, write, and watch files. Expose ports with public URLs. Full network access inside the sandbox.
Use cases
Let your AI agents write, run, and iterate on code in isolated environments.
Spin up Python sandboxes with pandas, numpy, and matplotlib pre-installed.
Run builds and tests in clean, reproducible environments every time.
Give users a real coding environment in the browser. No Docker required.